Chatbots and Aircargo Maturity
I discovered chatbots a year ago while I was exploring the potential of artificial intelligence doing the course deeplearning.ai on Coursera. Playing around a bit with Google technology, Dialogflow, I was mesmerized by the ease with which you can build a chat interface that generalizes what human beings intend to ask; and I immediately added a chatbot to my machine learning portfolio.
More or less half a year later, I joined CHAMP and started to look for new opportunities to innovate within our portfolio. Chatbots (or Conversational User Interfaces) quickly established themselves as a thing to explore, especially having known about past experiences with Alexa. However, the idea was to push the concept further, shifting from a simple digital assistant to something that could be offered as a new product or transforming the usage of an existing one.
Leveraging the opportunity to participate in the TIACA's Air Cargo Forum 2018 in Toronto, we decided to virtualize a booking agent to automate the quote request process. As usual, the intent behind this digitization was not to remove human beings but to free them from 80% of "business as usual," to enable them to focus more on the high-value aspects of their mission.
So, I quickly developed a proof of concept experiment, and we worked on modelizing the workflow in a chatbot framework. I will pass here the consideration regarding the production and technological challenges as I will write a dedicated article in the future for each of those aspects.
The idea was to develop the previous Alexa concept a bit further and provide a phone number that a customer could call to get a quote. A chatbot would answer him on the phone and gather the minimum information required to get a quote: origin, destination, the weight of the goods, and the date of the shipment. Of course, all of this gathering of data must be done a conversational manner, meaning that you should find the experience as pleasant as if speaking with a real human being. When the minimum amount of data has been collected, a request is made to the Champ next-gen quote module to get a set of options to give the customer with potential dates, flights, and cost of transportation. Furthermore, the user can choose the option with which he prefers to register the quote in the system. In addition to this workflow, we also designed a screen that can display the users the conversation in real time.
I had the opportunity to perform 38 demonstrations in Toronto. Of those 38, only four suffered a minor glitch, due to my strong French accent. One failed due to maintenance on our natural language processing provider. However, what is interesting is the variety of reactions that I got from the audience, who varied for the most positives ones with comments like:
"I love chatbots, we are operating in the B2C sector, and we put chatbots everywhere to speak with our customers. Right now, I have a huge team of people just on the phone asking everywhere where our shipments are. If I could have robots to make the calls, it would be invaluable."
To more conservatives:
"There is no way that my customers will ever want to chat with a robot, here on my continent we do not like those usages, and we prefer to interact with real agents."
Overall, I got most of the different kinds of reactions from my audience with a majority of curiosity and willingness to engage in a little experimentation. From my experience, this is a pretty good sign that we are in front of real innovation (there is no consensus around the idea), and secondly, the industry is keen to proceed further with it.
On my side, the result is obvious; I believe that the conversational user interface is the logical next step in user interaction: we evolved from green-on-blacks terminals to Windows-based GUI using keyboard and mouse. Then, eight years ago, we switched to touch interfaces on smart devices. Nowadays, the Z generation does not type their messages; they dictate them into their smartphones. The B2C, tracked by the appetite of the new generation for a highly personalized digital experience, will lead the path of this interaction revolution. However, the B2B world will be able to harvest a great set of tools to help ease and solve many of the issues it is facing and will face in the coming years.